Quantifying rill erosion processes in evolving soil landscapes: An experimental approach

Rill erosion is a hydrological-geomorphic process occurring on hillslopes across small spatial extents. Despite advances in soil erosion research, the dynamic feedbacks among the rill erosion process, morphological adaptation, and hydraulic characteristics remain poorly quantified. The current study introduces a high-resolution photogrammetric technology to resolve real-time coupling of rill network evolution with hydraulic characteristics and sediment yield. Four sets of successive rainfall simulation experiments were done with four rainfall intensities of 30, 60, 90, and 120 mm/h. A photogrammetric observation system was applied to dynamically capture the landscape evolutions of rill erosion during ongoing rainfall. The results showed that rainfall of 120 mm/h caused the most damage to soil surfaces with the soil loss being about 9 times higher than that at 30 mm/h. The key morphological indicators of rill density, splitting degree ( ω ), fractal dimension, and geomorphological comentropy ( H ) increased with rainfall duration and intensity, while the width-depth ratio decreased. H and ω were the parameters most closely related to sediment yield, contributing 61% and 39%, respectively. Based on H and ω , a comprehensive quantitative indicator (CQI) was established by regression analysis. The CQI increased exponentially with the rainfall duration. Among all hydraulic variables, the stream power was the best parameter to characterize the dynamic mechanism of soil erosion and the morphology. These findings are useful for understanding the rill erosion process on a slope and assessing landscape degradation by erosion.